Prevalence and Causes of Visual Impairment in Chinese Adults in Urban Southern China: The Liwan Eye Study

Objective  To assess the prevalence and causes of visual impairment and blindness in adults living in an urban area of southern China.

Methods  Random cluster sampling was used to identify the adults 50 years and older living in the Liwan district of Guangzhou, China. Presenting visual acuity (PVA) with habitual correction and best-corrected visual acuity (BCVA) based on autorefraction and subjective refraction were measured using the Early Treatment Diabetic Retinopathy Study visual chart. Blindness and low vision were defined according to World Health Organization criteria. Eyes with visual impairment were assigned 1 principal cause for the impairment.

Results  Visual acuity measurements were available for 1399 adults 50 years and older (75.3% participation rate). The prevalence of blindness and low vision based on the PVA was 0.6% (95% confidence interval, 0.2%-1.0%) and 10.1% (95% confidence interval, 8.5%-11.7%), respectively. These rates were reduced to 0.5% and 3.1% when the BCVA was considered. Based on the PVA, the principal causes for blindness were cataract (39.6%), glaucoma (11.0%), and myopic maculopathy (6.6%). The majority of low vision cases were attributable to cataract (45.3%) and uncorrected refractive error (43.9%).

Conclusion  The majority of eye diseases leading to visual impairment are potentially treatable in this population.

Visual impairment is recognized as an important social burden worldwide. The Vision 2020 project, a global initiative, was launched in 1999 with an aim to eliminate avoidable blindness by 2020.¹ Accurate information on the prevalence and causes of visual impairment may help international health organizations prioritize resources and develop appropriate human resources and infrastructures. Such information would help in the development of screening programs and the identification of people with an increased risk for diseases.

In 2002, the World Health Organization (WHO) estimated that there were 37 million people in the world with blindness and 124 million with low vision. The majority of these people live in developing countries. In China, the WHO estimated the rate of blindness in people 50 years and older as 2.3%.² This estimation was mainly based on 2 major population-based studies conducted in Shunyi,³ a suburban area near Beijing, and Doumen,⁴ a county in rural Guangdong Province. Both of these studies used the same protocol for sampling, examination, and disease definition. The results are widely recognized as being representative and accurate. However, these studies were conducted approximately 10 years ago, and the socioeconomic and demographic features of the Chinese population have changed in this time. Therefore, more up-to-date information with regard to visual impairment is needed.

In China, studies on the prevalence of visual impairment have been conducted by different groups of researchers.⁵ However, most of the data are derived from the samples with unknown representativeness. Because of the different measurement methods and nonuniform definitions used, the results are generally difficult to compare. Probably because they were more pragmatically feasible, most of the studies were conducted in rural areas of China.^3,4,6,7 Few studies were conducted for the urban population, where the health accessibility, socioeconomic status, and nutrition might be different. The Beijing Eye Study reports the prevalence of visual impairment in Beijing^8,9 but the study subjects were identified without random sampling. Furthermore, the situation in urban southern China remains largely unknown.

This study was motivated by an interest in obtaining visual impairment data in a representative urban setting in southern China. We aim to compare these data with the data obtained from other population-based studies in urban northern China and other rural areas of China. The comparative findings will provide new insight into the magnitude and changing pattern of visual impairment in Chinese adults living in urban settings as urbanization continues in China.

Guangzhou, the capital city of Guangdong Province, is the economic, cultural, and scientific center of southern China, with a population of 9.9 million in the 2000 census.¹⁰ The Liwan district, 1 of 10 administrative districts in Guangzhou, was identified for the survey because of its relatively stable population and representative demographic and socioeconomic characteristics. All the residents are of Han ethnicity (Chinese) and represent a wide socioeconomic spectrum. The average annual income per working resident is 14 198 yuan ($1715).¹¹

Detailed study procedures have been reported elsewhere.¹² In brief, adults 50 years and older were enrolled from the Liwan District of Guangzhou, using random cluster sampling. Fengyuan Street, with a population of 62 815 and occupying 0.77 km², was selected because it is a residential area with a limited number of commercial and industrial buildings. Ten clusters within this street block were identified. The clusters were defined geographically by the Residence Administrative Committees subdivisions, each with approximately 6000 residents. Those who had resided in the selected study sampling clusters for 6 months or longer were considered eligible during the enumeration. The eligible subjects were identified by their address, name of the household head, name of the subjects, and date of birth using the Household Resident Registry kept by the district government. Ethical approval was obtained from the Zhongshan University Ethics Review Board and the Research Governance Committee of Moorfields Eye Hospital in London, England. The study was conducted in accordance with the tenets of the World Medical Association Declaration of Helsinki. Written informed consent was obtained from all subjects. Examinations were performed between September 2003 and February 2004.

An examination station was set up in the community within half an hour's walking distance for most subjects. The identities of the subjects were verified using the subjects' official photo identity cards. Noncycloplegic refraction was measured using a handheld autorefractor (ARK-30; Nidek Corp, Gamagori, Japan). Distance visual acuity (VA) was measured using an Early Treatment Diabetic Retinopathy Study logMAR E chart (Precision Vision, Villa Park, Illinois) with a standard illumination box. Visual acuity measurement began at a distance of 4 m with the top line (20/200). If the orientation of at least 4 of the 5 optotypes was correctly identified, the subject was then tested by dropping down to line 4 (20/100). If 1 or fewer optotypes was missed, the testing continued at line 7 (20/50), to line 10 (20/25), and finally to line 11 (20/20). If at any level the subject failed to recognize 4 of the 5 optotypes, the line immediately above the failed line was tested, until the subjects successfully recognized the optotypes. If the subject failed to read the top line at 4 m, the subject was advanced to 2 m and then to 1 m, progressing down the chart as described earlier. The lowest line read successfully was assigned as the VA for the eye. For those with VA worse than 20/400, finger counting, hand motions, and no light perception were recorded along with examination distance. The presenting VA (PVA) with habitual refractive correction was recorded first and then the best-corrected VA (BCVA) was determined using the autorefraction result after the necessary refinement for those with a PVA less than 20/40 in either eye.

A comprehensive examination of each subject's eyelids, globes, lenses, and fundi was carried out using a slitlamp (Topcon SL-8Z [Topcon, Tokyo, Japan] with a Nikon D1 × digital image system [Nikon, Tokyo]) and a +78-diopter lens at ×16 magnification by an experienced ophthalmologist (M.H.). The pupils of patients with a BCVA of 20/40 or less were dilated. Similarly, the pupils of patients in whom lens and fundus status could not be evaluated satisfactorily were also dilated. If the PVA was less than 20/40 in any eye, 1 principal cause of impairment or blindness was assigned by the same experienced specialist (W.H.) using a 15-item diagnostic checklist according to the definitions in the study protocol. Refractive error (or uncorrected refractive error) was assigned if a patient presented with a VA of 20/40 or less that improved to more than 20/40 with best correction. Cataract was assumed for those with significant lens opacity that obscured observation of the fundus. Myopic maculopathy was assigned if the spherical equivalent was greater than −6.0 diopters and if significant damage (hemorrhage or atrophy) was observed on the retina and macula. Age-related macular degeneration (AMD) was assigned as a cause for visual impairment if significant damage was observed on the macula, such as exudative macular degeneration or geographic atrophy. Glaucoma was defined according to the International Society for Geographical and Epidemiological Ophthalmology Grading System.¹³ Other causes for visual impairment were determined according to clinical routine diagnosis. When 2 or more disorders might have caused the visual impairment in the same eye, the one most amenable to treatment or prevention was selected based on WHO recommendations.

To facilitate comparison with our previous findings in rural southern China, we used the same definition for visual impairment, by which VA in both eyes was taken into account: (1) normal/near normal, 20/63 or more in both eyes; (2) visual impairment, less than 20/63 and 20/200 or more in worse eye, 20/200 or more in better eye; (3) unilateral blindness, less than 20/200 in worse eye and 20/200 or more in better eye; (4) moderate bilateral blindness, less than 20/200 in worse eye, less than 20/200 and 20/400 or more in better eye; and (5) severe bilateral blindness, less than 20/400 in both eyes. The severe bilateral blindness category was equivalent to the WHO definition of blindness, while a combination of moderate bilateral blindness and severe bilateral blindness was equivalent to the US criterion for blindness. The visual impairment category was slightly different than either the WHO or US definitions because unilateral low vision was included, whereas the low vision category in both the WHO and US definitions only includes bilateral low vision. Therefore, the prevalence data were also presented based on WHO and US definitions for blindness and low vision.

Prevalence estimates with 95% confidence intervals (95% CIs) were calculated for the 3 blindness categories and visual impairment. Association of blindness with age, sex, and education was estimated using logistic regression.

Among the 1864 subjects eligible for the study, 1405 subjects 50 years and older were successfully examined, representing 75.3% of those eligible. Comparing the demographic characteristics of the enumerated sample with the total population of the Liwan district (2000 census), the subjects aged 60 to 69 years were underrepresented in the sample (27.4% in the sample vs 35.0% in the district). Otherwise, these 2 populations were similar. The demographic characteristics of participants and nonparticipants have been reported elsewhere.¹² In brief, men aged 50 to 59 years (63.6%) and people aged 80 to 93 years (59.5% for men and 60.4% for women) were less likely to participate, while the participation rates in other age groups were approximately 80%. Visual acuity data were not available in 6 subjects who were unfit for examination, leaving 1399 people available for analysis. The mean age was 65.3 years (range, 50-93 years), and 789 (56.4%) subjects were women.

The prevalence of blindness and low vision based on the PVA and BCVA is shown in Table 1. The crude prevalence of bilateral blindness (VA<20/200 in both eyes) was 1.2% (n = 16; 95% CI, 0.6%-1.7%) based on the PVA. When refractive errors were fully corrected, the rate of blindness decreased to 0.6% (n = 9; 95% CI, 0.3%-1.2%). Unilateral blindness (<20/200 in 1 eye) was found in 6.7% (n = 94) of the subjects (95% CI, 5.4%-8.0%) based on the PVA and dropped to 5.3% (n = 74) when the BCVA was considered. About 75.9% of people (n = 1062) were categorized as having normal or near normal vision (20/63 or better in both eyes). This percentage increased to 88.0% (n = 1231) when the BCVA was considered.

The association of unilateral and bilateral blindness (VA<20/200 in either eye) with age, sex, and education is summarized in Table 2. The prevalence of blindness increased with age. The adjusted odds ratio of blindness was 2.8 (95% CI, 1.5-5.3; P < .01) in people aged 60 to 69 years, 3.6 (95% CI, 2.0-6.6; P < .001) in those aged 70 to 79 years, and 5.9 (95% CI, 2.9-12.3; P < .001) in those 80 years and older when those aged 50 to 59 years were considered as the reference. The prevalence of blindness was not correlated significantly with sex (adjusted odds ratio, 1.1; P = .68). After adjusting for the effect of age and sex, those with a lower level of education (primary school and illiterate) tended to be associated with blindness.

Principal causes of visual impairment based on PVA and BCVA categories and age are summarized in Table 3. Among 486 eyes with a VA less than 20/63, 291 (59.9%) were found in women and 195 (40.1%) in men. Cataract was generally the most common cause of visual impairment across all levels of severity. However, the most frequent cause of visual impairment was cataract in the older individuals (≥70 years) while refractive error was more common in the younger group. Cataract and refractive error together accounted for approximately 90.0% of the eyes with a PVA less than 20/63 to 20/200 or more. In total, the principal causes of blindness (PVA<20/200 in both eyes) were cataract (38.9%), followed by refractive error (13.5%), glaucoma (8.7%), high myopic retinopathy (7.9%), AMD (6.3%), and retinitis pigmentosa (4.0%). Using BCVA as the criterion, refractive error was no longer a cause for visual impairment, as was expected, while the proportion of cataract, on the other hand, increased.

This study reports on the prevalence of visual impairment in adults living in urban southern China. The age-specific prevalence of low vision and blindness was substantially lower than what has been found in rural populations (Table 4). Consistent with studies of rural populations,^3,4 cataract was identified as the leading cause of low vision and blindness.

The population-based sampling design avoided the potential for selection bias commonly seen in hospital-based studies. Similar to most other population-based surveys, there was a slightly lower response rate in the younger men (aged 50-59 years) and elderly people (80 years and older) in the current study. The younger people who were reluctant to attend may have been those with better visual function while the older nonparticipants were more likely those with poorer health and more likely to have more severe visual problems. If this were the case, the prevalence of visual impairment might be somewhat overestimated in the younger people and biased downward in the older group. Given the reasonably high response rate, this issue is unlikely to affect our findings significantly.

The prevalence of low vision and blindness appears to vary across different regions in China and other countries. Table 4 summarizes age-specific rates for blindness and low vision in the surveys. The estimated prevalence of visual impairment in the current study, conducted in urban southern China, is much lower than the ones reported in the southern rural population (Doumen) and northern suburban population (Shunyi).^3,4 The prevalence was found to be even lower in both the urban and rural parts of the Beijing Eye Study.^8,9 Interestingly, in their cohort, they did not identify the urban/rural differences in visual impairment after controlling the confounders. There is no explanation for this, but perhaps the subjects in the Beijing Eye Study were not identified by random sampling. For this reason, the results may only apply to a selected group of people living in the defined areas.

The urban/rural difference in the rate of visual impairment appears to support directing more resources to rural areas. However, it does not indicate that resources should not be allocated in urban areas. To appreciate the implications of our data, one must consider them in the context of the total population in China. If we extrapolate the age- and sex-specific rate to the Chinese population in 2003 based on the US Census Bureau Web site,¹⁴ and an estimate of the urban to rural population ratio (urban to rural in 2003: 0.4053:0.5947) reported by the National Bureau of Statistics of China,¹⁵ the number of urban-dwelling adults (≥50 years) with visual impairment and blindness could be approximately 1.3 million and 0.9 million, respectively. These numbers highlight the need for prevention programs and visual impairment rehabilitation therapy both in urban and rural China.

There is increasing evidence suggesting an association between age, sex, socioeconomic status, and the risk of visual impairment.¹⁶ As expected, age was the primary predictor of blindness and visual impairment in this population. This result generally reflects the increasing threat posed by eye diseases over the next several decades as the Chinese population becomes older. Consistent with the findings of the Beijing Eye Study, the current study identified no sex difference in the magnitude of visual impairment and blindness.⁸ In addition, people with low levels of education were associated with visual loss. This is probably because of limited accessibility to eye care services and a lack of awareness of eye diseases among these people.

Our data confirm previous findings that the principal cause of blindness and low vision in China is cataract. Therefore, most visual impairment in this urban region was curable and/or reversible, representing a potential to improve the visual function of the adults living in urban countries. We also identified an increase in the proportion of cataract as a cause of blindness in this population with age, from 2.0% in people aged 50 to 59 years and 27.5% in those aged 60 to 69 years to 48.8% in those aged 70 to 79 years. This coincides with the lowest cataract surgical coverage rate in the 70 to 79 years age group, as shown in the Shunyi and Doumen studies. In the aforementioned study, the findings do suggest less optimal eye care accessibility in the elderly population, especially with regard to those 70 years and older. Cataract as a primary cause of blindness is not only found in developing countries, but also in relatively well-developed areas such as Hong Kong,¹⁷ Taiwan,¹⁸ and Singapore,¹⁹ where cataract surgical programs are believed to be widely available. Thus, besides the economy and health accessibility, other barriers, such as cultural perception and individual demand of vision, should also be considered.

Unlike industrialized Western countries,²⁰ where AMD is the predominant cause of blindness, there were only 8 cases (6.3%) of individuals identified as having AMD in those with a PVA less than 20/200 in the current study. This difference is probably partly due to genetic and/or environmental factors influencing the clinical presentation of AMD.²¹

In the current study, glaucoma was responsible for 8.7% (11 of 126 eyes) of blindness in the population (VA<20/200). This proportion is somewhat lower than in Singapore (60%)²² and Mongolia (35%),²³ where glaucoma was identified as the leading cause of blindness. The reason for this discrepancy is unclear. Our results, however, are similar to those reported in the Shunyi (45 of 611; 7.5%),³ Doumen (11 of 231; 4.8%),⁴ and Beijing (WHO blindness criteria, 7.7%)^8,24 studies. We recognize that visual field testing was not a criterion for blindness in our diagnostic protocols; this is because we intended to use similar criteria to our previous studies in Doumen and Shunyi. Some glaucoma cases may present with normal central vision but severe glaucomatous visual field defects may not have been detected. Because of this, glaucomatous blindness might have been underestimated. Because the majority of patients with glaucoma may not be aware of visual problems,^12,25 an effective screening program may be able to identify the cases that are in the early phases and prevent most patients with glaucoma from going blind.

Refractive errors have been recognized as the leading cause of visual impairment in industrialized nations²⁶ and the major cause of blindness in developing countries, such as India.²⁷ In the current study, refractive error was identified as the second principal cause, ranked after cataract, for visual impairment when PVA was used as the criterion for visual impairment. This was also found in the Shunyi and Doumen studies. When BCVA was used as the criterion, refractive error was no longer the cause for visual impairment. This is consistent with the findings in Beijing⁸ and Nantong,²⁸ where only the causes based on BCVA were reported. Recently, the WHO pointed out that using BCVA as the criterion might overlook the uncorrected refractive error. For this reason, the WHO recommends using PVA as the criterion for visual impairment.²⁹

In conclusion, our data estimate the prevalence of visual impairment and blindness in the Liwan district, an urban region in southern China. We confirmed previous findings that cataract and refractive error are responsible for a significant portion of blindness and visual impairment in this population. Avoidable and curable blindness is common in urban China. Age and low levels of education, but not sex, are significantly associated with increased rates of visual impairment and blindness. These findings justify the need for eye care programs and visual impairment rehabilitation services in urban regions. The social, psychological, and economic barriers that prevent these people from accessing eye care should be further explored.

Correspondence: Mingguang He, MD, PhD, Department of Preventive Ophthalmology, Zhongshan Ophthalmic Center, Guangzhou 510060, People's Republic of China (mingguang_he@yahoo.com).

Submitted for Publication: September 6, 2008; final revision received January 23, 2009; accepted January 30, 2009.

Financial Disclosure: None reported.

Funding/Support: Dr He is recipient of a University College London Graduate Research Scholarship and Overseas Research Scholarship (2001061054), grant 2005B30901008 from the Scientific and Technology Foundation of Guangdong Province, and Sun Yat-sen University Clinical Research 5010 Project. Dr Foster receives support from The Medical Research Council (grant G0401527), Wellcome Trust (grant 075110), and The Richard Desmond Charitable Foundation (via Fight for Sight).